Description of medical images in characteristic subspace and vector quantization coding based on wavelet transformation

With the novel idea of describing an image in characteristic subspace (DCS), a new approach of a more efficient image compression for some types of images, such as medical sciagraph, is proposed. Firstly, we obtain the coefficient image of the original image by its description in the DCS; then we decompose the coefficient image with a forward wavelet transform (WT) with multi-resolution to remove the redundant information; and finally encode the decomposed image by vector quantization (VQ). In comparison with WT-VQ or DCT-VQ, the new method (DCS-WT-VQ) has been proved by experiments to have a much higher compression ratio and SNR, requiring a lower computational time in the coding and decoding process.